Thread by @TehGimp

This is an interesting poll, and like all the other more-recent polls, it has some wrinkles that deserve to be unpacked.
#ableg #abvote #abpoli

https://twitter.com/markusoff/status/1113418232126300160

The published document includes the questionnaire, the cross-tabs, etc, which is super-awesome because it helps shed some light on why different polls might yield different results. Often these details are hidden behind a paywall.
#ableg #abvote #abpoli

https://twitter.com/planetjanetyyc/status/1113416317715881986

I believe this is a commissioned poll (for the Global Petroleum Show?) as well, like the recent EKOS poll (which was commissioned for Unifor). These tend to get excluded from poll-aggregators for a few good reasons:
#ableg #abvote #abpoli

https://twitter.com/EricGrenierCBC/status/1112767915584049154

That's why the CBC poll tracker includes the recent Research Co poll, but not that EKOS poll and probably not this poll either.
#ableg #abvote #abpoli

The basics: the poll was conducted Mar 25-30 (so of course does not include reaction to Mark Smith, etc) and surveyed 900 people which is a decent sample. Lots of fun details about the minutiae of how they reached out to people, how they could respond, etc.
#ableg #abvote #abpoli

The leadership approval numbers are perhaps the least interesting--they jive really well with other polls that also show a modest edge for Notley in overall approval. The best predictors of approval ratings (p4) are all too familiar:
#ableg #abvote #abpoli

No demo surprises in party support either (p5). Most polls that report these factors affirm that education is the best predictor of NDP support, & being old or living outside #yeg/#yyc are the best predictors of UCP support. Take from that what you will.
#ableg #abvote #abpoli

The included data tables (p10 onward) naturally confirm these details. Across several polls, education seems to be the best overall predictor of party support. Again, take from that what you will.
#ableg #abvote #abpoli

The meat of the poll is what's interesting. This poll seems to show no major vote intention changes since the writ was dropped, esp in #yyc, whereas other recent polls show the UCP struggling to maintain support (though still safely in majority territory).
#ableg #abvote #abpoli

All polls have skewed demographics compared to the actual voting population. Some pollsters will weight or exclude responses based on demographics, in hopes of making their results more reflective of actual voters. This poll doesn't seem to do that AFAICT.
#ableg #abvote #abpoli

If you're a progressive looking for a reason to not worry about this poll (which I would not recommend), then your best bet is probably the income breakdown (p69-71). 1/3 of respondents make >$120000/year, which is not reflective of AB as a whole.
#ableg #abvote #abpoli

I had trouble pinning down a good source, but given StatsCan's numbers, the >$120000 crowd probably make up somewhere in the neighbourhood of <10% of voters, certainly not 1/3rd.
#ableg #abvote #abpoli
www12.statcan.gc.ca/census-recense…

i.e. this poll appears to skew heavily towards reflecting the opinions of the wealthy, those lucky few that actually pay more tax under the NDP. Unless the responses were weighted accordingly, this could this poll a lot less representative than some others.
#ableg #abvote #abpoli

Other demographic factors seem to be more or less in agreement with the general voting pop, at least at a glance. High-education voters appear to be slightly over-represented, for example, but the skew isn't as huge as with income.
#ableg #abvote #abpoli

And that's all this particular non-expert has to say about this poll at the moment. Polls are polls, and should be treated accordingly--the only one that matters is on April 16.
#ableg #abvote #abpoli

A clarification on who exactly commissioned this particular poll:
#ableg #abvote #abpoli

https://twitter.com/peterloliver/status/1113442495910670342